Re: Only one compilation unit.

[Kilgallen@eisner.decus.org.nospam (Larry Kilgallen)]

In article <9f6e2b77.0107160541.10707819@posting.google.com>, B_Gaffney@My-Deja.com (B.Gaffney) writes:
> Kilgallen@eisner.decus.org.nospam (Larry Kilgallen) wrote in message news:<1luFIcjdO+HI@eisner.encompasserve.org>...
>> In article <x7vu20gzlxq.fsf@smaug.pushface.org>, Simon Wright <simon@pushface.org> writes:
>> > John English <je@brighton.ac.uk> writes:
>> > 
> ...
>> > We had fun when the source was called Foo.ads, under NT one of the
>> > split files ended up also called Foo.ads -- which overwrote our
>> > original. Also, GNAT does *not* like files with upper-case letters in
>> > the name at all!
>> 
>> Even on VMS ?
> 
> I don't believe so, I think this is specific to Windows.
> 
> I had a problem with this and GnatMake told me to submit it as a bug,
> which I did.  The reply I got was that GNAT was working correctly.
> 
> If I had to guess (with fear of being corrected), I'd say that, while
> Win32 allow mixed case, it is case-insensitive, and some of the tools
> GNATMAKE uses are UNIX-based and case-sensitive.  This probably
> wouldn't cause a problem on an OS that either didn't care about case
> (i.e. VMS) or was case-sensitive.

As I see it, the VMS situation should be comparable to that on Windows
when the VMS system is using ODS-5 disks, which are case-insensitive but
case-preserving.